GitHub

Designed to retrieve data from GitHub, this loader parses Markdown (.md) and text (.txt) files only. It can also auto-refresh and track any changes to these files, notifying the user of missing or new content. Advanced features will include code identification and hyperlink parsing.

If you need to load documentation in any other format (such as .mdx), use the Website agent to load the documentation from the published website.

Configuration

The GitHub agent can be configured to track specific information about a linked repository.

You can only have one agent per branch in a repository. However, you are free to set up agents on different branches in the same repository.

Configuration option	Explanation
Repository url	URL of GitHub repository to get data from
Repository branch	Branch of GitHub repository to get data from
File types	Types of files to pull when searching through this repository. These can only be markdown (.md) and/or text (.txt) files. If you need to load documentation in any other format (such as .mdx), use the Website agent to load the documentation from the published website.
Schedule	When to automatically run this agent
Include PRs	If enabled, the agent will also include pull requests in the data it tracks. We track the PR's related issue, changes to any files, requested reviewers, and when it was merged if it already has been merged.
Include Issues	If enabled, the agent will also include issues in the data it tracks. We track the content of the issues as well as relevant metadata.
Tags	Tags are applied to all future data collected by this agent, which can be used later for filtering.